Int. J. Environ. Res. Public Health 2014, 11, 6246-6264; doi:10.3390/ijerph110606246 OPEN ACCESS
International Journal of Environmental Research and Public Health ISSN 1660-4601 www.mdpi.com/journal/ijerph Article
The Deployment of Carbon Monoxide Wireless Sensor Network (CO-WSN) for Ambient Air Monitoring Chaichana Chaiwatpongsakorn, Mingming Lu, Tim C. Keener * and Soon-Jai Khang Department of Biomedical, Chemical, and Environmental Engineering, University of Cincinnati, 2901 Woodside Dr., Cincinnati, OH 45221, USA; E-Mails:
[email protected] (C.C.);
[email protected] (M.L.);
[email protected] (S.-J.K.) * Author to whom correspondence should be addressed; E-Mail:
[email protected]; Tel.: +1-513-556-3676; Fax: +1-513-556-4162. Received: 24 April 2014; in revised form: 3 June 2014 / Accepted: 4 June 2014 / Published: 16 June 2014
Abstract: Wireless sensor networks are becoming increasingly important as an alternative solution for environment monitoring because they can reduce cost and complexity. Also, they can improve reliability and data availability in places where traditional monitoring methods are difficult to site. In this study, a carbon monoxide wireless sensor network (CO-WSN) was developed to measure carbon monoxide concentrations at a major traffic intersection near the University of Cincinnati main campus. The system has been deployed over two weeks during Fall 2010, and Summer 2011–2012, traffic data was also recorded by using a manual traffic counter and a video camcorder to characterize vehicles at the intersection 24 h, particularly, during the morning and evening peak hour periods. According to the field test results, the 1 hr-average CO concentrations were found to range from 0.1–1.0 ppm which is lower than the National Ambient Air Quality Standards (NAAQS) 35 ppm on a one-hour averaging period. During rush hour periods, the traffic volume at the intersection varied from 2,067 to 3,076 vehicles per hour with 97% being passenger vehicles. Furthermore, the traffic volume based on a 1-h average showed a good correlation (R2 = 0.87) with the 1-h average CO-WSN concentrations for morning and evening peak time periods whereas CO-WSN results provided a moderate correlation (R2 = 0.42) with 24 hours traffic volume due to fluctuated changes of meteorological conditions. It is concluded that the performance and the reliability of wireless ambient air monitoring networks can be used as an alternative method for real time air monitoring.
Int. J. Environ. Res. Public Health 2014, 11
6247
Keywords: carbon monoxide; wireless sensor network; monitoring; traffic; intersection
1. Introduction Carbon monoxide (CO) is an odorless, tasteless and colorless gas that can be found at dangerous concentrations both indoors and outdoors [1]. Ambient CO comes primarily from automobile exhaust and high concentrations have been reported in enclosed garages, along roadways, and near intersections. In 2011, U.S. EPA reported that on-road and non-road vehicles contribute approximately 41 million short tons which accounts for 50% of the national carbon monoxide emissions [2]. CO can directly affect the health of people who work or live nearby these areas. After inhalation into the respiratory system, it eventually prohibits hemoglobin (Hb) in blood cells binding and carrying oxygen molecules because it reacts with hemoglobin faster than oxygen does [3]. Inhaling at high concentration of CO can result in dizziness, headaches, unconsciousness, and even death [4]. Therefore U.S. EPA has established the National Ambient Air Quality Standards (NAAQS) for CO in 1971. Then, U.S. EPA decided to retain the current standard at 9 ppm for 8-h average and 35 ppm for 1-h average after the most recent review of NAAQS for CO in 2011 [5]. Recently many studies have been accomplished to scrutinize the correlation between CO exposure and traffic. Kaur et al. [6] have studied the personal exposure and the carbon monoxide at the street canyon intersections. The results indicated that the personal CO exposure levels were high during morning rush hours. Chan et al. [7] have indicated that the CO concentrations from vehicular exhausts contributed more CO levels to nearby pedestrians than to people who are in vehicles. As a result of the studies above, and an increasing community concern to CO exposure due to the implementation of more bicycle lanes, more pedestrians and more mixed traffic as a consequence of the driving-less initiatives, it is necessary to have enough 24-h CO monitoring stations covering potential hot-spot areas. In order to measure CO concentrations continuously, U.S. EPA recommends non-dispersive infrared (NDIR) as a standard reference method. The principle of this technique is based on absorption of infrared radiation by CO molecules in the wavelength of 4.7 µm region [8]. This method has several advantages. For example, its measurement is generally not affected by an ambient temperature, and it provides a good sensitivity to the broad concentration range with a short response time. However, this method requires a substantial infrastructure including calibration gases, pumps, monitoring stations and other peripheral equipment along with a rather expensive analyzer [9]. Because of the high expense, these measurements can only be conducted at very limited sites. For instance, in the United States, there were approximately 328 CO monitoring stations as of May 2011 but only 52 CO monitoring sites are operating near roadways [5]. In addition, many previous studies have shown that the pollutant measurement from fixed monitoring stations provide underestimated measurements compared with the actual exposure levels [10,11]. This error was caused by the variation in sampling heights, the distance from the hot spot area, and the insufficient sampling rate. Recent advances in wireless sensor networks (WSNs) have shown an alternative solution for monitoring ambient air quality. Some WSN systems have been developed for short-term and long-term operations. For instance, Chung and Oh [12] developed a wireless sensor to monitor the indoor CO2
Int. J. Environ. Res. Public Health 2014, 11
6248
concentrations. A significant advancement of this study lies in that the circuit module has various enlargements for installing other sensors such as humidity, CO2, and flying dust sensors, etc. Furthermore, Cordova-Lopez et al. [13] have integrated a Geography Information System (GIS) with a wireless sensor network using CO, CO2, NO, NO2, SO2, and hydrocarbon sensors for monitoring tailpipe emissions from vehicle exhaust. They found that, during idle conditions when the car is waiting for a red traffic light, hydrocarbons and CO2 were the lowest compared with other pollutants while at a speed of 70 km/h nitrogen oxides (NOx) levels reached the highest concentration 100 ppm,. However, no ambient evaluation has been performed for monitoring carbon monoxide with WSNs. In this study, a CO-WSN system is comprised of a carbon monoxide sensor, a humidity sensor, and a temperature sensor integrated with a data acquisition board, a wireless communication system and a solar panel for monitoring CO concentrations continuously without charging. According to the preliminary experimental results, the lowest detection limit of the CO sensor is 0.1 ppm which is equivalent to the U.S. EPA reference method. Sensor will provide the advantage of flexibility in a deployment, a lower operation and a maintenance cost compared with the NDIR reference method, and will offer much more real time information about the spatial and temporal variations of ambient CO concentrations. This type of information is crucial for use by the transportation planners, environmental regulators, and the general public. 2. Materials & Methods 2.1. Components of CO-WSN A CO-WSN is comprised of five functional units: an analog sensor unit, a data acquisition board, a wireless radio module, a solar panel unit, and a radio antenna (Figure 1). Figure 1. Components of the CO wireless sensor network (CO-WSN).
Int. J. Environ. Res. Public Health 2014, 11
6249
2.1.1. Analog Sensor Unit Each analog sensor unit consists of a CO sensor and a signal amplifying board. There are three different options for the CO sensor: chemo-optical, semiconductor and electro-chemical. Among them, the electro-chemical sensors are typically found to report more accurate CO concentrations and are inexpensive in comparison with the others. Electro-chemical sensors generally use an acid as electrolyte and platinum (Pt) as a catalyst to react with the carbon monoxide gas and give off electrons. The principle chemical reactions inside the sensor are shown as follow [14–16]: At working electrode: CO + H2O CO2 + 2H+ + 2e− At counter electrode: ½O2 + 2H+ + 2e− H2O Overall reaction: CO + ½O2 CO2
(1) (2) (3)
At the working electrode, carbon monoxide reacts with water and provides electrons in proportion to the CO concentration while oxygen reacts with hydrogen ions and electrons and gives off water as a product at the counter electrode. The current generated at the working electrode, usually reported as voltages, can be used to determine CO concentration in the ambient. In our system design, the CO sensor connected to the data acquisition board (MDA 300CA) that can convert an analog signal of the CO sensor to a digital signal and then, transmitted to the MicaZ base station (MIB520) by a MicaZ radio transceiver module. Furthermore, the temperature and relative humidity (RH) can be detected by sensors that are located on the data acquisition board (MDA 300CA). In this study a three electrode electrochemical CO sensor model RCO100F (KWJ Engineering Inc., Newark, CA, USA) was customized and assembled into the wireless network system. The sensor specification is also shown in Table 1. Table 1. CO sensor model RCO100F specifications. Characteristics Size (W × L × H) Measuring Range Measuring Principle Onboard Filter Output Signal, Zero, 25 °C Output Signal, Span, 25 °C Lower Detection Limit Resolution Repeatability Output Linearity Response Time (t-90) Long Term Drift—Span Operating Temperature Range
Specifications 3.2 × 3.2 × 2.0 cm 0–100 ppm Electrochemical Oxidation of CO To remove SOx, NOx & H2S
